{"id":"W3027597167","doi":"","title":"Daydream: Accurately Estimating the Efficacy of Performance Optimizations for DNN Training","year":2020,"lang":"en","type":"article","venue":"USENIX Annual Technical Conference","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Profiling (computer programming); Artificial neural network; Machine learning; Software; Artificial intelligence; Graph; Parallel computing; Computer engineering; Theoretical computer science; Programming language","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001650941,0.001985225,0.0005041473,0.00145174,0.0004561126,0.0008561694,0.001906527,0.0009651612,0.00149077],"category_scores_gemma":[0.01358725,0.0008444602,0.0005924713,0.0008149632,0.0007312308,0.002549624,0.0007324674,0.00207673,0.0007843673],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001883873,"about_ca_system_score_gemma":0.002252705,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01507407,"about_ca_topic_score_gemma":0.02827667,"domain_scores_codex":[0.9987978,0.0002503411,0.00007997356,0.0003559358,0.000373277,0.0001425993],"domain_scores_gemma":[0.9949666,0.002872327,0.0004100608,0.0009097323,0.0006743135,0.0001669783],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007540266,0.0003554248,0.03757213,0.0004185147,0.0002177717,0.0001369999,0.0001745481,0.7604238,0.01540773,0.003129577,0.01429616,0.1671133],"study_design_scores_gemma":[0.00001655866,0.00007653696,0.002191417,0.00001393477,0.00001490271,0.00002334811,0.00001660363,0.9848934,0.01008938,0.001548049,0.001099082,0.00001683033],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6299272,0.002218967,0.3052942,0.0008084406,0.0003400398,0.0002270171,0.004869027,0.04836909,0.00794613],"genre_scores_gemma":[0.8325489,0.0004781463,0.1570958,0.0003371222,0.00003644396,0.0002083273,0.005180962,0.001862726,0.002251664],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01507407,"threshold_uncertainty_score":0.02997267,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1356333107934018,"score_gpt":0.3383242010037694,"score_spread":0.2026908902103675,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}